Olivia
Online
▲ Marktechpost AI
@Marktechpost
The fastest AI dev news engine on X — model releases, tools, and what they actually mean
What is trending in AI?
Joined April 2016
1.2K Following
11.9K Followers
14.1K Posts
▲ IBM just released Granite Speech 5.0 2B, an Apache 2.0 speech model that transcribes and translates speech in 1 model.
The key idea is a CTC Conformer encoder feeding a Granite 4.0 1B LLM through a q-former. Audio reaches the LLM at 10 Hz. Prompts switch between transcription, translation, language ID and keyword lists of up to 2,000 terms.
It ranks #2 among open-weight models on the Open ASR Leaderboard at 5.68% average English WER. Qwen3-ASR-1.7B leads at 5.40%. Cohere Transcribe sits at 5.77%. It is a trade-off though: on noisy FFASR audio it scores 14.80% vs Qwen3-ASR's 13.41%, and Japanese WER is 10.74%.
2B params, 6 ASR languages, Apache 2.0. Runs in Transformers, vLLM, llama.cpp and mlx-audio. IBM also updated Turbo CTC (19,870 RTFx, now 7 languages) and 4.1 2B Plus......
Full analysis: https://t.co/PQFyBSCoes
Model: https://t.co/rPMD6OPh5P
Blog: https://t.co/Ua1vyZlQoi
Paper: https://t.co/U0JPgTQ6OR
@IBM @IBMwatsonx @IBMData @IBMResearch
▲ Perplexity and NVIDIA just released JOLT, a post-training method where one model is both teacher and student, and the teacher is trained to teach.
The key idea is simple. Self-distillation gives the model a verified answer and lets it score its own attempts. But a teacher holding the answer can take shortcuts the student cannot follow. JOLT trains the teacher with outcome reward plus a KL pull toward the student, so its feedback stays within reach.
On MATH (Qwen3-0.6B), JOLT matched GRPO's accuracy with 6.5x fewer processed training tokens, and about 13x fewer completion tokens. JOLT+ (adds student reward) beat GRPO by 9.9 points on AppWorld and hit 10.96% pass@1 on Terminal-Bench 2.1 vs 8.25%. It is a trade-off though: without student reward, JOLT trails GRPO on AppWorld (51.92% vs 62.20%).
It is a method, tested on Qwen3 0.6B to 30B-A3B and DeepSeek-V2-Lite.
Full analysis: https://t.co/LxJXefu0tD
Paper: https://t.co/k4nb9CspuV
Blog: https://t.co/9Ahg4Ovl6Q
@perplexity_ai @AravSrinivas
@perplexity_ai https://t.co/lx3R78TJOZ
▲ Perplexity and NVIDIA just released JOLT, a post-training method where one model is both teacher and student, and the teacher is trained to teach.
The key idea is simple. Self-distillation gives the model a verified answer and lets it score its own attempts. But a teacher holding the answer can take shortcuts the student cannot follow. JOLT trains the teacher with outcome reward plus a KL pull toward the student, so its feedback stays within reach.
On MATH (Qwen3-0.6B), JOLT matched GRPO's accuracy with 6.5x fewer processed training tokens, and about 13x fewer completion tokens. JOLT+ (adds student reward) beat GRPO by 9.9 points on AppWorld and hit 10.96% pass@1 on Terminal-Bench 2.1 vs 8.25%. It is a trade-off though: without student reward, JOLT trails GRPO on AppWorld (51.92% vs 62.20%).
It is a method, tested on Qwen3 0.6B to 30B-A3B and DeepSeek-V2-Lite.
Full analysis: https://t.co/LxJXefu0tD
Paper: https://t.co/k4nb9CspuV
Blog: https://t.co/9Ahg4Ovl6Q
@perplexity_ai @AravSrinivas
▲ Perplexity and NVIDIA just released JOLT, a post-training method where one model is both teacher and student, and the teacher is trained to teach.
The key idea is simple. Self-distillation gives the model a verified answer and lets it score its own attempts. But a teacher holding the answer can take shortcuts the student cannot follow. JOLT trains the teacher with outcome reward plus a KL pull toward the student, so its feedback stays within reach.
On MATH (Qwen3-0.6B), JOLT matched GRPO's accuracy with 6.5x fewer processed training tokens, and about 13x fewer completion tokens. JOLT+ (adds student reward) beat GRPO by 9.9 points on AppWorld and hit 10.96% pass@1 on Terminal-Bench 2.1 vs 8.25%. It is a trade-off though: without student reward, JOLT trails GRPO on AppWorld (51.92% vs 62.20%).
It is a method, tested on Qwen3 0.6B to 30B-A3B and DeepSeek-V2-Lite.
Full analysis: https://t.co/LxJXefu0tD
Paper: https://t.co/k4nb9CspuV
Blog: https://t.co/9Ahg4Ovl6Q
@perplexity_ai @AravSrinivas
Who to follow
Salon Intake
@salonintake
AI-powered salon operations platform. Helping beauty professionals spend less time on paperwork and more time with clients. Invite only, coming soon.
Varun Yerram
@varunyer
PhD'ing @NYUDataScience | Ex @GoogleDeepMind | IITG
Zeistencroix
@zeistencroix
Post industrial noir // Members: OD & Santi https://t.co/I8no1f3God
@tydsh https://t.co/AjPgdpkyuy
▲ Recursive releases Flint, an Apache 2.0 local coding MoE with 40B total and ~2.5B active params (74.2% SWE-bench Verified)
The key idea is sparsity tuned for local hardware. Flint has 39.8B parameters but activates only ~2.5B per token: a router picks 6 of 256 experts plus 1 shared expert. Recursive's Eureka Machine, an automated research system, searched the architecture, the 4-bit quantization recipe and the kernels.
Recursive reports 74.2% on SWE-bench Verified vs 73.4% for Qwen3.6-35B-A3B, and 125+ tok/s on an M4 Pro MacBook Pro (150+ on a 64 GB RTX Spark). It is a trade-off though: on Terminal-Bench 2.1 Flint scores 54.9%, while dense Qwen 3.8 hits 73.0%.
Weights are Apache 2.0. 4-bit builds take about 23 GB and need 48 GB of memory. rec, Recursive's local engine, targets M4 Pro and M5 Pro/Max......
Full analysis: https://t.co/CTlQJWnSI1
Model: https://t.co/v5m6KqT3Fr
Repo: https://t.co/B946g2i6Tx
Blog: https://t.co/6btwU7ft2K
@Recursive_SI @RichardSocher @CaimingXiong @tydsh
@CaimingXiong https://t.co/AjPgdpkyuy
▲ Recursive releases Flint, an Apache 2.0 local coding MoE with 40B total and ~2.5B active params (74.2% SWE-bench Verified)
The key idea is sparsity tuned for local hardware. Flint has 39.8B parameters but activates only ~2.5B per token: a router picks 6 of 256 experts plus 1 shared expert. Recursive's Eureka Machine, an automated research system, searched the architecture, the 4-bit quantization recipe and the kernels.
Recursive reports 74.2% on SWE-bench Verified vs 73.4% for Qwen3.6-35B-A3B, and 125+ tok/s on an M4 Pro MacBook Pro (150+ on a 64 GB RTX Spark). It is a trade-off though: on Terminal-Bench 2.1 Flint scores 54.9%, while dense Qwen 3.8 hits 73.0%.
Weights are Apache 2.0. 4-bit builds take about 23 GB and need 48 GB of memory. rec, Recursive's local engine, targets M4 Pro and M5 Pro/Max......
Full analysis: https://t.co/CTlQJWnSI1
Model: https://t.co/v5m6KqT3Fr
Repo: https://t.co/B946g2i6Tx
Blog: https://t.co/6btwU7ft2K
@Recursive_SI @RichardSocher @CaimingXiong @tydsh
▲ Recursive releases Flint, an Apache 2.0 local coding MoE with 40B total and ~2.5B active params (74.2% SWE-bench Verified)
The key idea is sparsity tuned for local hardware. Flint has 39.8B parameters but activates only ~2.5B per token: a router picks 6 of 256 experts plus 1 shared expert. Recursive's Eureka Machine, an automated research system, searched the architecture, the 4-bit quantization recipe and the kernels.
Recursive reports 74.2% on SWE-bench Verified vs 73.4% for Qwen3.6-35B-A3B, and 125+ tok/s on an M4 Pro MacBook Pro (150+ on a 64 GB RTX Spark). It is a trade-off though: on Terminal-Bench 2.1 Flint scores 54.9%, while dense Qwen 3.8 hits 73.0%.
Weights are Apache 2.0. 4-bit builds take about 23 GB and need 48 GB of memory. rec, Recursive's local engine, targets M4 Pro and M5 Pro/Max......
Full analysis: https://t.co/CTlQJWnSI1
Model: https://t.co/v5m6KqT3Fr
Repo: https://t.co/B946g2i6Tx
Blog: https://t.co/6btwU7ft2K
@Recursive_SI @RichardSocher @CaimingXiong @tydsh
▲ Recursive releases Flint, an Apache 2.0 local coding MoE with 40B total and ~2.5B active params (74.2% SWE-bench Verified)
The key idea is sparsity tuned for local hardware. Flint has 39.8B parameters but activates only ~2.5B per token: a router picks 6 of 256 experts plus 1 shared expert. Recursive's Eureka Machine, an automated research system, searched the architecture, the 4-bit quantization recipe and the kernels.
Recursive reports 74.2% on SWE-bench Verified vs 73.4% for Qwen3.6-35B-A3B, and 125+ tok/s on an M4 Pro MacBook Pro (150+ on a 64 GB RTX Spark). It is a trade-off though: on Terminal-Bench 2.1 Flint scores 54.9%, while dense Qwen 3.8 hits 73.0%.
Weights are Apache 2.0. 4-bit builds take about 23 GB and need 48 GB of memory. rec, Recursive's local engine, targets M4 Pro and M5 Pro/Max......
Full analysis: https://t.co/CTlQJWnSI1
Model: https://t.co/v5m6KqT3Fr
Repo: https://t.co/B946g2i6Tx
Blog: https://t.co/6btwU7ft2K
@Recursive_SI @RichardSocher @CaimingXiong @tydsh
@vercel https://t.co/Z66EGbJPxG
@GuillaumeLample https://t.co/EutdQE0g4J
▲ Mistral AI just released Voxtral 3, a full voice stack: 3 transcription models, 1 TTS model and a voice agent SDK.
The key idea is 1 streaming pipeline. Realtime Transcribe 3 listens and detects end of turn. An LLM like Mistral Small 4 reasons and calls tools. Voxtral Speech 3 replies in the conversation's tone, with each turn under 1 second.
Voxtral Small Transcribe 3 scores 2.1% AA-WER, #3 on Artificial Analysis, ahead of Scribe v2 (2.2%). Diarization hits 18.4% cpWER vs 20.3% for MAI-Transcribe-2. It is a trade-off though: MAI-Transcribe-2 is about 4.7x faster, and the realtime model at 240 ms scores 7.4% WER vs 4.2% for Gemini 3.5 Realtime.
Public preview on the Mistral API (EU, US, global). Batch ASR from $0.0035/min, TTS $13 per 1M characters, voice agent API from $0.04/min. No open weights announced......
Full analysis: https://t.co/4YLhfaFkEW
Announcement: https://t.co/kB6aeWRCgJ
Docs: https://t.co/0XF6GtabKt
@MistralAI @arthurmensch
▲ Mistral AI just released Voxtral 3, a full voice stack: 3 transcription models, 1 TTS model and a voice agent SDK.
The key idea is 1 streaming pipeline. Realtime Transcribe 3 listens and detects end of turn. An LLM like Mistral Small 4 reasons and calls tools. Voxtral Speech 3 replies in the conversation's tone, with each turn under 1 second.
Voxtral Small Transcribe 3 scores 2.1% AA-WER, #3 on Artificial Analysis, ahead of Scribe v2 (2.2%). Diarization hits 18.4% cpWER vs 20.3% for MAI-Transcribe-2. It is a trade-off though: MAI-Transcribe-2 is about 4.7x faster, and the realtime model at 240 ms scores 7.4% WER vs 4.2% for Gemini 3.5 Realtime.
Public preview on the Mistral API (EU, US, global). Batch ASR from $0.0035/min, TTS $13 per 1M characters, voice agent API from $0.04/min. No open weights announced......
Full analysis: https://t.co/4YLhfaFkEW
Announcement: https://t.co/kB6aeWRCgJ
Docs: https://t.co/0XF6GtabKt
@MistralAI @arthurmensch
▲ Google, Red Hat and IBM Devs just released llm-d 1.0, turning a fleet of vLLM and SGLang servers into one coordinated inference pool on Kubernetes.
The key idea is cache-aware routing. The router tracks which replica already holds each agent's prefix and sends the next step there. When GPU or TPU memory fills, KV cache spills to host RAM, SSD or Managed Lustre instead of being evicted.
Google reports Gemini's prefix cache hit rate doubled from 35% to 70%, and a 17,000-node GKE benchmark scaled throughput 2.9x. It is a trade-off though: every number is self-reported, and RL time-slicing still leaves roughly 30% of accelerator time idle.
Apache 2.0, CNCF Sandbox, runs on NVIDIA, AMD, Intel GPUs and Google TPUs......
Full analysis: https://t.co/NSi0TgLQuv
Repo: https://t.co/pc81smywtL
Docs: https://t.co/Oj7dMDTD0A
Technical details: https://t.co/7JGfhsE2S4
@GoogleAI @googleaidevs @RedHat @RedHat_AI @IBM @IBMData @IBMResearch
@MistralAI https://t.co/EutdQE0g4J
▲ Mistral AI just released Voxtral 3, a full voice stack: 3 transcription models, 1 TTS model and a voice agent SDK.
The key idea is 1 streaming pipeline. Realtime Transcribe 3 listens and detects end of turn. An LLM like Mistral Small 4 reasons and calls tools. Voxtral Speech 3 replies in the conversation's tone, with each turn under 1 second.
Voxtral Small Transcribe 3 scores 2.1% AA-WER, #3 on Artificial Analysis, ahead of Scribe v2 (2.2%). Diarization hits 18.4% cpWER vs 20.3% for MAI-Transcribe-2. It is a trade-off though: MAI-Transcribe-2 is about 4.7x faster, and the realtime model at 240 ms scores 7.4% WER vs 4.2% for Gemini 3.5 Realtime.
Public preview on the Mistral API (EU, US, global). Batch ASR from $0.0035/min, TTS $13 per 1M characters, voice agent API from $0.04/min. No open weights announced......
Full analysis: https://t.co/4YLhfaFkEW
Announcement: https://t.co/kB6aeWRCgJ
Docs: https://t.co/0XF6GtabKt
@MistralAI @arthurmensch
@arthurmensch https://t.co/EutdQE0g4J
▲ Mistral AI just released Voxtral 3, a full voice stack: 3 transcription models, 1 TTS model and a voice agent SDK.
The key idea is 1 streaming pipeline. Realtime Transcribe 3 listens and detects end of turn. An LLM like Mistral Small 4 reasons and calls tools. Voxtral Speech 3 replies in the conversation's tone, with each turn under 1 second.
Voxtral Small Transcribe 3 scores 2.1% AA-WER, #3 on Artificial Analysis, ahead of Scribe v2 (2.2%). Diarization hits 18.4% cpWER vs 20.3% for MAI-Transcribe-2. It is a trade-off though: MAI-Transcribe-2 is about 4.7x faster, and the realtime model at 240 ms scores 7.4% WER vs 4.2% for Gemini 3.5 Realtime.
Public preview on the Mistral API (EU, US, global). Batch ASR from $0.0035/min, TTS $13 per 1M characters, voice agent API from $0.04/min. No open weights announced......
Full analysis: https://t.co/4YLhfaFkEW
Announcement: https://t.co/kB6aeWRCgJ
Docs: https://t.co/0XF6GtabKt
@MistralAI @arthurmensch
▲ Mistral AI just released Voxtral 3, a full voice stack: 3 transcription models, 1 TTS model and a voice agent SDK.
The key idea is 1 streaming pipeline. Realtime Transcribe 3 listens and detects end of turn. An LLM like Mistral Small 4 reasons and calls tools. Voxtral Speech 3 replies in the conversation's tone, with each turn under 1 second.
Voxtral Small Transcribe 3 scores 2.1% AA-WER, #3 on Artificial Analysis, ahead of Scribe v2 (2.2%). Diarization hits 18.4% cpWER vs 20.3% for MAI-Transcribe-2. It is a trade-off though: MAI-Transcribe-2 is about 4.7x faster, and the realtime model at 240 ms scores 7.4% WER vs 4.2% for Gemini 3.5 Realtime.
Public preview on the Mistral API (EU, US, global). Batch ASR from $0.0035/min, TTS $13 per 1M characters, voice agent API from $0.04/min. No open weights announced......
Full analysis: https://t.co/4YLhfaFkEW
Announcement: https://t.co/kB6aeWRCgJ
Docs: https://t.co/0XF6GtabKt
@MistralAI @arthurmensch
▲ Google, Red Hat and IBM Devs just released llm-d 1.0, turning a fleet of vLLM and SGLang servers into one coordinated inference pool on Kubernetes.
The key idea is cache-aware routing. The router tracks which replica already holds each agent's prefix and sends the next step there. When GPU or TPU memory fills, KV cache spills to host RAM, SSD or Managed Lustre instead of being evicted.
Google reports Gemini's prefix cache hit rate doubled from 35% to 70%, and a 17,000-node GKE benchmark scaled throughput 2.9x. It is a trade-off though: every number is self-reported, and RL time-slicing still leaves roughly 30% of accelerator time idle.
Apache 2.0, CNCF Sandbox, runs on NVIDIA, AMD, Intel GPUs and Google TPUs......
Full analysis: https://t.co/NSi0TgLQuv
Repo: https://t.co/pc81smywtL
Docs: https://t.co/Oj7dMDTD0A
Technical details: https://t.co/7JGfhsE2S4
@GoogleAI @googleaidevs @RedHat @RedHat_AI @IBM @IBMData @IBMResearch
▲ DeepSeek just open-sourced DeepInfer, the inference engine that serves DeepSeek-V4.1-Flash in production.
The key idea is DSpark: a draft model proposes 5 tokens, and a Confidence Head predicts which ones will be accepted. The scheduler verifies only the most valuable prefix. Add Wide-EP decode, SWA bounded replay and async scheduling, and you get DeepSeek's real serving stack.
The numbers: above 25,000 decode tokens/s per GPU on 64 GPUs (EP64, 131K context). DSpark verifies about 30% fewer draft tokens with accepted length nearly flat (3.29 vs 3.20). It is a trade-off though: this release is offline only, with no KV cache store, P/D disaggregation or load balancer, and it supports 1 model.
Apache 2.0. Runs on NVIDIA SM100 GPUs and Huawei Ascend 950 NPUs. deepseek-recipe (MIT) ships alongside......
Full analysis: https://t.co/XSiuQtk5Fd
Repo: https://t.co/kkSKp8Ab9h
Docs: https://t.co/VMAsYDxEzM
Model: https://t.co/pdviLOA2qF
@deepseek_ai
▲ Google just open-sourced TorchTPU, a native PyTorch backend that makes Google TPUs an ordinary PyTorch device.
The key idea is PrivateUse1 registration. Tensors stay plain PyTorch tensors on torch.device("tpu"), with no JAX runtime. Code runs eager by default, then torch.compile lowers it through StableHLO to XLA.
The headline number: TorchTitan reached 52% MFU training Qwen 3 8B, with near-linear scaling up to 8,192 Ironwood TPUs. It is a trade-off though: out-of-box eager took 11.49s per step in Google's fine-tuning case, versus 2.78s after compile, kernels and sharding.
TorchTPU is co-developed with Meta, Apache 2.0, v0.2.0 on PyPI, and replaces PyTorch/XLA. vLLM, SGLang and Hugging Face Transformers v5 already run on it......
Full analysis: https://t.co/cGTAmk6YQG
Repo: https://t.co/yUKljkhgHH
Paper: https://t.co/yU8OAwaYu6
Docs: https://t.co/BgRcyvIAnm
@googlecloud @GoogleAI
▲ Google just open-sourced TorchTPU, a native PyTorch backend that makes Google TPUs an ordinary PyTorch device.
The key idea is PrivateUse1 registration. Tensors stay plain PyTorch tensors on torch.device("tpu"), with no JAX runtime. Code runs eager by default, then torch.compile lowers it through StableHLO to XLA.
The headline number: TorchTitan reached 52% MFU training Qwen 3 8B, with near-linear scaling up to 8,192 Ironwood TPUs. It is a trade-off though: out-of-box eager took 11.49s per step in Google's fine-tuning case, versus 2.78s after compile, kernels and sharding.
TorchTPU is co-developed with Meta, Apache 2.0, v0.2.0 on PyPI, and replaces PyTorch/XLA. vLLM, SGLang and Hugging Face Transformers v5 already run on it......
Full analysis: https://t.co/cGTAmk6YQG
Repo: https://t.co/yUKljkhgHH
Paper: https://t.co/yU8OAwaYu6
Docs: https://t.co/BgRcyvIAnm
@googlecloud @GoogleAI
Last Seen Users on Sotwe
PublicSexHub
Seen from Turkey
Türk İfşa Perisi
Seen from Turkey
Telegram Türk İfşa +18
Seen from Belgium
Geniş Grup
Seen from Turkey
video colmek indo
Seen from Indonesia
Kaleel🇱🇾🇸🇾
yacine soumis kelb
Seen from Algeria
Khong Biet
Seen from Korea
yarrak delisi😍😍
Seen from Turkey
Lal Us
Seen from Netherlands
Most Popular Users
Elon Musk
@elonmusk
241.5M followers
Barack Obama
@barackobama
118.8M followers
Cristiano Ronaldo
@cristiano
114.5M followers
Donald J. Trump
@realdonaldtrump
111.8M followers
Narendra Modi
@narendramodi
107.1M followers
Rihanna
@rihanna
98.7M followers
NASA
@nasa
92.3M followers
Justin Bieber
@justinbieber
91.8M followers
KATY PERRY
@katyperry
90M followers
Taylor Swift
@taylorswift13
84M followers
Lady Gaga
@ladygaga
75.5M followers
Virat Kohli
@imvkohli
73.5M followers
Kim Kardashian
@kimkardashian
70.9M followers
YouTube
@youtube
68.8M followers
Neymar Jr
@neymarjr
66.5M followers
Bill Gates
@billgates
65.3M followers
Selena Gomez
@selenagomez
63.1M followers
The Ellen Show
@theellenshow
62.2M followers
CNN
@cnn
61.7M followers
X
@x
60.7M followers